Cognitive Electronic Jamming Decision-Making Method Based on Improved -Learning Algorithm

Pseudocode 1

Input: jamming action set , working state set , initialization matrix , reward matrix , discount factor and other parameters.

Output: optimal jamming strategy.

1: Begin.

2: While // is the maximum number of iterations.

3: For // is the number of restart cycles.

4: Calculate the restart period according to equation (9);

5: For .

6: Update the learning rate according to equation (6);

7: Update the temperature according to equation (4);

8: Randomly initialize the working state;

9: While (the current state is not the target working state).

10: Randomly select action for the current state from the jamming action set;

11: Choose optimal action according to equation (2);

12: Calculate the exploration probability according to equation (3);

13: Generate random number between [0,1];

14: If

15: action=;

16: Else

17: action=;

18: End if.

19: Execute the current action, update the radar state, and obtain the.

reward value according to equation (10);

20: Update the function according to equation (1);

21: Calculate the difference ;

22: If // is the convergence threshold.

23: Jump out of the loop and terminate the learning process;

24: End if

25: End while

26: End for

27: Update the learning rate range according to equations (7) and (8), and restart the learning rate;

28: // is the initial restart period.

29: End for

30: End while

31: Output table to get the optimal interference strategy;

32: End.